Research of data mining and network coverage optimization in early warning model of chlorine gas monitoring wireless sensor network

Massive historical data are stored in chlorine gas monitoring network. So that prediction algorithm of data mining is used to dig historical data not only can make the redundant data reused, but also can forecast the network trend and improve the network early warning model. The chlorine gas monitoring wireless sensor network based on ZigBee was designed in this paper. Then Fletcher-Reeves algorithm was added to dig historical data in the network, forecast the network trend and improve the early warning model. To avoid network monitoring blind areas and save the number of monitoring nodes, the coverage optimal mathematical model of chlorine monitoring wireless sensor network is established. The optimal matching of the node number and the network coverage is realized by using artificial fish-swarm algorithm. The predicted concentration of chlorine data were trained by data mining model. The maximum relative error between predicted concentration and measured concentration was 11.08%, and the maximum average error was 7.36%. And it can satisfy actual requirements.